No Brother print drivers? y0wza, that's bad. Brother is maybe the only printer brand worth a darn anymore.
Apologies if it was linked here as well and I missed it:
Weird that M1 is listed at 3GHz frequency. Wasn’t it 3.2?This gives this (I hope I did not make errors...)
View attachment 101577
So Oryon is below x3 in int work/GHz.
My 16 inch M1 Pro laptop is looking better by the day. So glad I picked it up a year ago with a discount.Can no one beat Firestorm?
No one can do it, not even the person who created it, lol. What a beast of an architecture!!
Fours years and still kicking
M1 should be at 3.2GHz.I mean those numbers are funny. M1 -> M2 is an IPC regression??
That's entirely possible. Even the same uarch that's clocked higher can score lower per clock in a memory bandwidth-bound scenario as additional clock cycles aren't doing anything.View attachment 101580
I did this math, and I discovered that Firestorm in M1 has higher PPC than Firestorm in A14 Bionic, and even Avalanche in A15 Bionic.
Yeah I guess the Baidu guy made a mistake. With 3.2 GHz the M1 score/GHz goes from 2.47 to 2.31, placing it behind M2 as I would expect.
View attachment 101582
Yeah that seems dubious. Modern CPU frequency variations are a real pain for that kind of cross CPU uarch comparisons. I guess the only acceptable way would be to lock CPU frequency the closest possible to boost frequency, but I'm afraid that's not possible for all platforms.Unfortunately I think there are multiple errors for the frequencies. 7700x has a 5.4 GHz listed max boost, 13700k is also 5.4 GHz. Technically the 7700x has an Fmax of 5.55 GHz but it's unlikely to actually run that high in SPEC. For sure it's not getting 5.65 GHz. M2 clocks at 3.5 GHz (maybe 3.49 technically), not 3.45 GHz. I believe the X3 goes up to 3.36 GHz as well but someone could check me on that.
These results look a little suspect to me but I guess it's the best we have for now, but for sure, the listed frequencies are not reliable.
As far as "IPC" goes, this is not massively different from what we have above.Stop testing exclusively desktop Intel/AMD chips on Spec for christ’s sake. Anyways, here’s another datapoint with MTL, Raptor Mobile, Phoenix, M3.
View attachment 101600
It's possible some people don't take the terms of the license very seriously.Any clue as to why it's mostly Chinese sources that are SPEC licensees? We have Geekerwan, David Huang and now whoever this is.
That's the mythical 4.3 GHz devkit I guess.Have you guys forgotten this claim?
View attachment 101605
Not that.That's the mythical 4.3 GHz devkit I guess.
Power is of course unrealistic, but the 1T score is likely realistic. It's just that the slide was done before QCOM got RDNA3'd.Not that.
From the numbers we have now current X Elite SKUs, there is no way the 4.3 GHz SKU is going to match the M2 Max at 30% less power.
This has been posted above and there are some doubts about the frequencies in the graph.Somebody ran SPEC2017 on X Elite: